"Ice Box" the apparently not-quite-as-efficient-as-it-should-be ice maker cooling method.

[Blazing Falken]

Re-titled since this isn't worth the effort. 3 consecutive days of sleep deprivation can do wonders. Sorry, folks.

 

Nothing super fancy this time but possibly the most efficient way to abuse game mechanics, this time courtesy of the ice maker.

Full speed operation will require one dupe dedicated to operating it but the more the merrier.

Materials are aluminum and ceramic, coolant loop is polluted H2O. The main pool is filled with a total of 2,000 kg of water. Vacuum in lower chamber is recommended for best efficiency.

Doors are optional depending on how much cold bleed you find acceptable vs. power/time lost. Note that in later screenshots there is a door to the lower pump area. This door is NOT optional as it is for recovering dropped water bottles... because dupes are jerks like that.

The 2 metal tiles to the right are where you would load heat to be removed. The 2x2 water pocket to the right is being used for demonstration purposes.

Updated conveyor screenshot to match original design. Bridge direction AND position were off. Sleep is important.

At first the dupes will be idle frequently because water takes a long time to cool and freeze. This phase will not last long. As the temperature of the water pool dives, so does the "slow" portion of the cooling process. Once the water inside the ice maker goes below -0.6c, the water becomes ice. Since ice has a higher conductivity than water, the ice maker cools it down to -20c much more rapidly. Since the ice maker is still generating heat at a fixed rate, the heat removal has greater efficiency.

Pre-heated room  and all contents to ~50c for demonstration.

Pool temps after just 3 cycles:

After 10 cycles:

After 15 cycles: We've reached nearly max efficiency where water being fed to the ice makers is almost instantly turned to ice and reaches -20c very quickly.

After 20 cycles:

External 4,000kg pool heat removed after 20 cycles (36c starting temp, 4.2c final.) Keep in mind, this is all while removing nearly 50c of heat from the entire system in the process. Cooling ability from this point onward will be a lot stronger.

 

 

 

[Promethien]

I love it. Just a quick thought. Instead of having the side pool there, why not run the ice through the water pool directly then use a radiant pipe through the water to carry the cool where you want it to go?

[WanderingKid]

I like it, but how efficient is it per ice maker?  I've been tinkering with ice from the ice biome that I needed to dig out for pathing, an ice maker, an automated dispersal unit, and an ethanol bath.  Using a pump to pull water that floats above the ethanol to keep it colder, all for a 'cheap' bean farm.  Curious if it's just easier to go this route.

[Lifegrow]

Lovely  

Have you considered running a conveyor loop through your cooling pool for transferring temperatures?

You could easily slap a thermo sensor in the pool leading to a conveyor shutoff. That way you could cool materials providing your cooling pool was cold enough, and shut it off if it warmed too much.

Great stuff though bud, still havn't built an ice maker yet  

[Promethien]

3 hours ago, WanderingKid said:

I like it, but how efficient is it per ice maker?  I've been tinkering with ice from the ice biome that I needed to dig out for pathing, an ice maker, an automated dispersal unit, and an ethanol bath.  Using a pump to pull water that floats above the ethanol to keep it colder, all for a 'cheap' bean farm.  Curious if it's just easier to go this route.

Once the water is riding the freezing point its around twice as efficient as the original calculations for ice makers people were making back when they were last changed.

The OP made it no secret that this is very dupe labor intensive. The demonstrated room will keep a dupe occupied nearly full time. By my calculation pre-super coolant this is cheaper on power than an aquatuner + turbine.

[Blazing Falken]

If you run the conveyor through the side pool water directly, then it will melt into the heat load and be both out of the pump range and way hotter, ruining efficiency. Remember, the side pool is for demonstration more or less.

The pool in the bottom needs to be isolated from any heat in order to maintain the near -0.6c temperature. The higher the incoming water temperature is to the ice makers, the more inefficient they become. Any time spent cooling the water within the ice maker above -0.6c is creating more heat.  This is the most important aspect to the design. Water has a conductivity of 0.609 while ice has 2.180. That's ~3.58x faster (I probably messed the numbers up, it seems more like 2x-ish.. I just woke up though so.... correct me if you please)  bringing the already frozen ice to -20c.

If the conveyor rail is run through the pool directly, then the water will freeze and can't be pumped.

This is why the "output" cooling tiles remains outside of the system that cools and maintains the inner workings.

As the ice runs through the metal tiles it cools them off. When it melts, it drops back into the pumping pool at near freezing.

[Promethien]

Ah right I was noticing the efficiency but missed the step you took there to ensure it stays in the ideal range. 

[Blazing Falken]

We want to keep the room cooling as first priority and the output cooling separated and secondary as much as possible. Obviously if the room is too hot the dupe(s) will suffer scalding injury plus the risk of equipment damage.

The ice on the conveyor itself does all the work and can not transfer heat back into the system this way.

As for dupe labor, I suggest having a single dupe isolated with only access to living needs and the ice box.This keeps other dupes from using the pump for other purposes. Don't worry, your dupe will be happy to exist playing such a vital role. (I.e. don't bother with moral boosting or skill assignments.)

[mathmanican]

8 hours ago, Blazing Falken said:

Since ice has a higher conductivity than water, the ice maker cools it down to -20c much more rapidly. Since the ice maker is still generating heat at a fixed rate, the heat removal has greater efficiency.

Thinking out loud here, and would love corrections to my thoughts before I build this. 

The SHC of water is higher than the ice, which explains the more rapid cooling once you hit the ice phase. The heat generated is the same, when it is ice or water, because it's removing the same amount of heat, hence ice cools faster?  Either way, you see about 8.8 kDTU/s of cooling (going from 95C to -20 C, or from 0C down to -20C). So the heat deletion capabilities of the icemaker don't matter based on how you build something....  With lower SHC, the ice temp will drop really fast (conductivity doesn't really matter here - it only matters when you take the ice out and start melting it). 

I've been thinking about this post most of the day, and not seeing how this will be any better than cooling 95C water, except with one caveat. If the game does not cause the ice to jump up in temp when it phase changes (as happens outside the machine), then the extra cooling we might be  seeing may be from when the ice reverts to water, and drops in temp (getting free cooling from the phase change back to water). Is this what's going on? 

[Blazing Falken]

@mathmanican You may be right on the SHC being the mechanic here, makes sense that the observed cooling rate is roughly doubled as the SHC is roughly halved.

Also, the temperature of the 30kg water does not increase when it changes into ice inside the machine. You can watch it go from -0.6c water to -0.7c ice without any other change, aside from the more rapid cooling rate.

[KittenIsAGeek]

28 minutes ago, mathmanican said:

I've been thinking about this post most of the day, and not seeing how this will be any better than cooling 95C water, except with one caveat. If the game does not cause the ice to jump up in temp when it phase changes (as happens outside the machine), then the extra cooling we might be  seeing may be from when the ice reverts to water, and drops in temp (getting free cooling from the phase change back to water). Is this what's going on? 

I've had similar thoughts.  I suspect the phase change has a lot to do with how well these work.  For example, an ice-E fan doesn't really accomplish much ... UNTIL you start getting a lot of cold bottles of water showing up.  Once that happens, these things can really cool a lot.

Here's my observations:

In your industrial area, you have an ice maker.  It produces -20c chunks of ice that dupes run out to the ice-E fans.  Some dupe jumps on and starts operating it and  the building works as the tooltip shows -- which is kinda pathetic.  Once the ice melts, it phase changes into a bottle of water.  That bottle of water can accept a LOT of thermal energy before reaching ambient temperatures.  So, lets pretend that your dupes aren't using bottled water for anything and that bottle stays there.  Well, the dupes grab another hunk of ice, throw it in the ice-E machine, and start pedaling... and soon there's a second bottle.  And a third.

Sounds somewhat familiar, right? *cough*algae terrariums*cough*.  Anyway, eventually you get a LOT of very cold water sitting there absorbing heat from the surrounding area.

Lets assume that the phase change happens right at 0c and the bottle appears at 1c.  I know this isn't quite how it works, but its close enough for now.  Lets also assume that your desired ambient is 20c.  The ice delivered to your Ice-E fans comes in 10kg chunks.  When the ice transitions to a bottle, we go from 10kg * 2.05 DTU/^oC to 10kg * 4.179 DTU/^oC.  The ice, then, has -410k DTU of energy, while the water (1^oC warmer), has -794k DTU.  If the hystresis is 10^oC difference between ice and water, then you get -418k DTU -- which is still a decrease of 8k DTU in the overall thermal energy of the mass.  In other words, once melted, water can take a hella-lotta (technical term) thermal energy before reaching ambient as compared to ice (assuming it doesn't melt).  

 

[mathmanican]

10 minutes ago, Blazing Falken said:

Also, the temperature of the 30kg water does not increase when it changes into ice inside the machine. You can watch it go from -0.6c water to -0.7c ice without any other change, aside from the more rapid cooling rate.

So we are abusing SHC mechanics again.   Beautiful. The crying crab all over.  Anything that can dodge the phase change will freely generate gobs of extra cooling (or heating).  I'll do some experiments, though you can probably beat me to it. 

This may be worth a bug report.

[KittenIsAGeek]

1 minute ago, mathmanican said:

So we are abusing SHC mechanics again. 

We do it IRL, why not ONI?  Of course, IRL we use it to move heat into the (relatively) infinite sink of the environment and in ONI we straight-up destroy it, but.. hey, the end result is all we care about.  Right?  Did we enjoy the abuse?  My 16c apartment suggests I do.  =^.^=

[Blazing Falken]

4 hours ago, mathmanican said:

So we are abusing SHC mechanics again.   Beautiful. The crying crab all over.  Anything that can dodge the phase change will freely generate gobs of extra cooling (or heating).  I'll do some experiments, though you can probably beat me to it. 

Looks like it! I haven't had much sleep over the past few days so my assumptions are a bit hazy, but the results speak volumes still.

Feel free to further experiment with this, I have a new idea** on abusing internal phase change for cooling in a most unusual way. Off to the sandbox I go...

The pursuit of alternative cooling methods after we lost the water sieve temperature clamping has proven to be very fruitful.

**Edit: Mother of god... I may have finally done it. Some more testing and refinement is needed but this might be huge... I can't say just yet but... let's hope I'm right for the sake of the experiment.  **wasn't repeatable.

[KittenIsAGeek]

My favorite real-world abuse of SHC is the cooling system on the Saturn V rocket that took astronauts to the moon.  A small reservoir of water feeds water through a sublimater where it is exposed to space.  The phase changes from liquid to solid to vapor dissipates 9kw of heat.

Watch at least from minute 3 to minute 5.  Though I highly recommend the entire video.

Anyway, it meant that the electronics could be kept at 10c to 15c for the entire flight using only a small reservoir of water.

[WanderingKid]

So I've been tinkering with a single ice maker cooling system in Sandbox.

My math seems horribly wrong, so if someone can check it?

After full stabilization, I added in enough heat into the heat sink (right side) after shutting off the ice maker system.  I then stabilized the heat sink to 85F (yes, F, you'll live).  The lower chamber is vacuum and water.

After a cycle, which will use the ice maker four times (120 kg of ice in the conveyor overall) the O2/Copper in the heat sink went from 85F to 70F, for a 15F change.

SHC of the 90kg of O2 is 0.558 DTU/g/F.  SHC of the 300 kg of copper is 0.214 DTU/g/F.

This gets me a total cooling of 963,000 DTU for the Copper and 753,300 DTU for the O2.  Totals out to 1,716,300.

Over ~600 seconds that's roughly 2,860.5 DTU, or 2.86 kDTU worth of cooling.  That number seems HORRIBLY wrong.  What did I miss?

[bobucles]

39 minutes ago, WanderingKid said:

Over ~600 seconds that's roughly 2,860.5 DTU, or 2.86 kDTU worth of cooling.  That number seems HORRIBLY wrong.  What did I miss?

That sounds about right. The ice maker generates 16kDTUs of heat. Supposedly that factors in the 20% heat deletion, so there should be 16-20 = 4kDTUs of cooling overall. The autosweeper generates 2kDTUs of heat when active, and the rest of the system likely has inefficiencies or heat losses as well. A single ice machine simply isn't enough to get good numbers from.

[KittenIsAGeek]

OK, you also need to consider the water.  First, the heat emitted from the ice maker includes the heat "taken" from the water to turn it into ice.  Second, ice has about half the SHC of water.  Your setup isolates the ice from the water, so you're only getting cooling until the ice melts.

[WanderingKid]

30 minutes ago, KittenIsAGeek said:

OK, you also need to consider the water.  First, the heat emitted from the ice maker includes the heat "taken" from the water to turn it into ice.  Second, ice has about half the SHC of water.  Your setup isolates the ice from the water, so you're only getting cooling until the ice melts.

Right, the heat transfer removes 20% of the heat from the system per system cycle.  Once my water was 'almost' frozen I was getting 20% deletion 4x per cycle.

If the cooling water is not isolated, than the ice maker cycle will take longer as it needs to cool longer. If there isn't a vacuum isolation for the cooling chamber, then you can't keep the lower water at 'nearly' freezing, as this was built for.  Assuming you're going to run this system to cool something... let's take Electrolizer O2 as an example.

Trying to cool 1kg/second O2 from 70C to 25C requires x cooling. You dump that heat into the cooling chamber and it eventually removes it .  If you also dump that heat into the water below, it will eventually stabilize higher than the 'almost ice' temperatures that the system prefers.

If it's just for fun and giggles and it's not meant to be an working and active cooling system, then I misunderstood the point.  Right now it appears a natural wheeze in hydrogen is more effective than the 60w of power, nevermind the rest of it.

[KittenIsAGeek]

I understand what you're trying to do.  That wasn't really my point. My point was that ice isn't a great coolant.  Water, on the other hand, is.  Yes, Ice is more conductive and can reach colder temperatures, but it can only accept half the thermal energy of water.  So, if you want the room to cool down, the water has to be part of your cooling process, rather than simply being fed into the input.

[WanderingKid]

Wouldn't it make more sense then to run the water to almost boiling THEN shove it into the ice maker?  That's your fastest heat removal by time... yes, I realize the scalding O2 isn't useful at that temp, but I'm trying to understand where there's any value in this long term.  Everything I'm understanding, even doubling the SHC cooling values of Water to ice, means a single fertilized Wheeze is worth ~2.5 ice makers in this setup.

I realize there's value in the phase change somewhere, but I just don't understand it.  It seems backwards to me.

[KittenIsAGeek]

Here's the value in the phase change:  Lets start by setting our goal at -20^oC.  1kg of water at 0^oC has 83.6k DTU of energy.  1kg of ice at 0^oC has 41k DTU of energy -- about half the energy of water.  The phase change, from water to ice, "destroys" 42.6k DTU of thermal energy.

[mathmanican]

On 8/31/2019 at 2:17 PM, WanderingKid said:

I realize there's value in the phase change somewhere

I played with this for hours last night.

• Water (outside the icemaker) turns to ice at 269.5K (checked with debug to 4 decimals).
• Ice jumps to 271K (estimate) instantly on phase change.
• Ice converts to water at 272.5K (esimate)
• Water appears at 272.5K (best estimate)

Note that there is NO change in temp when ice swaps back to water. This means that the ice maker's changing of water to ice (at 272.5K) does NOT create an SHC abuse. 

On the other hand, if you freeze water to ice, and then melt it, you generate heat.  Getting the water to freeze requires dropping the ice temp (hence increasing temp of surrounding atmosphere).  To melt the ice, you have to heat it (dropping temp of surroundings) about 1/4th as much as needed for freezing water.

You end up generating heat by rapidly phase changing from water to ice, and back to water, provided you do not use the ice maker.  Using the ice maker prevents this from occurring.  Question: could I abuse this to run several steam turbines.

Edit: Looks like @DonDegow  decided to do this.  

[KittenIsAGeek]

You should try piping the melted water through the radiator, emptying into a small reservoir for the pitcher pump to feed the ice maker.  Perhaps something like this:

You should see an overall cooling because of how the ice maker turns water to ice.

[WanderingKid]

11 minutes ago, KittenIsAGeek said:

Here's the value in the phase change:  Lets start by setting our goal at -20^oC.  1kg of water at 0^oC has 83.6k DTU of energy.  1kg of ice at 0^oC has 41k DTU of energy -- about half the energy of water.  The phase change, from water to ice, "destroys" 42.6k DTU of thermal energy.

Right, but you 'regenerate' that heat when you go back to water.  In this case, the system is neutral... but according to @mathmanican's comments above as well..

11 minutes ago, mathmanican said:

On the other hand, if you freeze water to ice, and then melt it, you generate heat.  Getting the water to freeze requires dropping the ice temp (hence increasing temp of surrounding atmosphere).  To melt the ice, you have to heat it (dropping temp of surroundings) about 1/4th as much as needed for freezing water.

You end up generating heat by rapidly phase changing from water to ice, and back to water, provided you do not use the ice maker.  Using the ice maker prevents this from occurring.

I'm going to have to read that a few times when I have more caffeine apparently.  What I heard you say was "1.5K of temperature is cheating the system."  No, it's not what you said, but it's what I comprehended.

Thank you both.